INDEX

Explanations

instances of the word "talk" in different contexts

oai_token-act-pair · gpt-3.5-turbo

instances of the word "talk" and its variations, suggesting a focus on communication

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

cffff

-0.71

rant

-0.68

emale

-0.65

rypt

-0.65

ynski

-0.65

Released

-0.65

rikes

-0.64

feeding

-0.63

arte

-0.62

lisher

-0.61

POSITIVE LOGITS

 about

1.14

 smack

0.97

about

0.96

 ABOUT

0.93

 aloud

0.85

 frankly

0.84

 nonsense

0.83

ative

0.83

 About

0.81

 loudly

0.81

Activations Density 0.045%

instances of the word "talk" in different contexts

instances of the word "talk" and its variations, suggesting a focus on communication

No Comments

No Known Activations